Abstract
The fungal pathogen Fusarium graminearum infect both heads and roots of cereal crops causing several economically important diseases such as head blight, seedling blight, crown rot and root rot. Trichothecene mycotoxins such as deoxynivalenol (DON), a well-known virulence factor, produced by F. graminearum (Fg) during disease development is also an important health concern. Although how F. graminearum infects above-ground tissues is relatively well studied, very little is known about molecular processes employed by the pathogen during below-ground infection. Also unknown is the role of DON during root infection. In the present study, we analyzed the transcriptome of F. graminearum during root infection of the model cereal Brachypodium distachyon. We also compared our Fg transcriptome data during root infection with those reported during wheat head infection. These analyses suggested that both shared and unique infection strategies employed by the pathogen during colonization of different host tissues. Several metabolite biosynthesis genes induced in F. graminearum during root infection could be linked to phytohormone production, implying that the pathogen likely interferes root specific defenses. In addition, to understand the role of DON in Fg root infection, we analyzed the transcriptome of the DON deficient Tri5 mutant. These analyses showed that the absence of DON had a significant effect on fungal transcriptional responses. Although DON was produced in infected roots, this mycotoxin did not act as a virulence factor during root infection. Our results reveal new mechanistic insights into the below-ground strategies employed by F. graminearum that may benefit the development of new genetic tools to combat this important cereal pathogen.
Introduction
Fungal plant pathogens have adopted versatile strategies to colonize their hosts. While some fungal pathogens show strict host and tissue specificity, others can adjust their lifestyles to infect different hosts and tissues. Some fungal pathogens such as the rice blast Magnaporthe oryzae (Marcel et al., 2010; Sesma and Osbourn, 2004) and the corn smut Ustilago maydis (Mazaheri-Naeini et al., 2015), which commonly invade above-ground plant parts, can also undergo developmental processes resembling to root infecting fungi. Such changes in the pathogen may require sensing of host signals and we previously showed that sensing of root signals prior to root infection by Fg can indeed lead to developmental changes in the pathogen (Ding et al., 2020). As a member of the Fusarium species complex, Fusarium graminearum (Fg) causes Fusarium Head Blight (FHB) or scab, one of the most economically important diseases of cereal crops. FHB causes substantial yield losses and mycotoxin contaminations of grains, resulting in billions of dollars of economic losses worldwide and threatening our food supply and safety (Chen et al., 2019; Trail, 2009). Most studies on FHB have so far focused on wheat heads as the pathogen initially infects individual wheat florets from which it can spread to other florets through the rachis and can eventually colonize the whole spike. However, recent research has shown that Fg is also capable of infecting roots and young seedlings of wheat, barley and maize, causing crown rot, root rot and seedling blight (Henkes et al., 2011; Lanoue et al., 2010; Stephens et al., 2008; Wang et al., 2015; Zhou et al., 2019). During the initiation of root infection, Fg forms a peg structure outside the root surface and move inter- and intra-cellularly without causing root necrosis. This early colonization stage is followed by a transition of the fungus to a necrotrophic life style where lesions develop and spread to stems and aboveground tissues (Wang et al., 2015).
The availability of the complete genome sequence of Fg makes investigations of global regulation of gene expression in this fungus feasible (Kazan and Gardiner, 2018a; Ma et al., 2013). Infection strategies of Fg evaluated by transcriptome analyses in different hosts and tissues exclusively during infection of their above-ground tissues such as heads, stems and coleoptiles revealed mostly distinct, but also common gene expression patterns (Boedi et al., 2016; Harris et al., 2016; Lysøe et al., 2011; Zhang et al., 2012, 2016). Interestingly, different cereal species can produce different defense-related metabolites (Dutartre et al., 2012) and such differences may explain why Fg might need to tailor its arsenal during colonization of different hosts (Harris et al., 2016). A recent comparative transcriptomic study of Fg also showed differential expression of fungal genes during infection of FHB resistant and susceptible wheat genotypes (Pan et al., 2018).
Although transcriptome studies have provided clues associated with host specificity of the Fg infection process, fungal transcriptomes of Fg mutants with altered virulence have rarely been tested on the same hosts. The mycotoxin deoxynivalenol (DON) is a Fg virulence factor during infection of wheat heads (Proctor et al., 1995). DON may also be needed during the interaction of Fg with its broader environment (Audenaert et al., 2013). The first step in DON biosynthesis is catalyzed by the trichodiene synthase Tri5, which cyclizes farnesyl pyrophosphate to trichodiene (Hohn and Beremand, 1989). In contrast to its strong expression pattern during wheat head infection, Tri5 did not show increased in planta expression during wheat coleoptile infection, suggesting that DON’s effect on pathogen virulence is tissue specific (Zhang et al., 2012). Other studies have suggested a crucial role for DON during the colonization of wheat stems by Fg and the related pathogens F. pseudograminearum and F. culmorum (Desmond et al., 2008; Mudge et al., 2006; Powell et al., 2017; Scherm et al., 2013a). The recent finding where Fhb7-mediated FHB and crown rot disease resistance relies on DON detoxification also highlighted virulence function of this mycotoxin in wheat (Wang et al., 2020). Furthermore, DON is known to activate defense gene expression in wheat (Desmond et al., 2008). Indeed, host transcriptional changes observed in Brachypodium distachyon (Bd) and wheat spikelets infected by the Tri5 deletion mutants (ΔTri5) differed from those by wildtype Fg (Brauer et al., 2020; Pasquet et al., 2014). DON biosynthesis in Fg is regulated by Tri6 and Tri10 transcription factors. Analyses of deletion mutants for these genes by transcriptome profiling during plant infection revealed significant transcriptional alterations for a large number of genes, many of which have not been implicated previously in toxin production (Seong et al., 2009). Genetic analyses undertaken in Fg have identified many genes influencing DON biosynthesis (Chen et al., 2019). However, DON-non-producing mutants have not been employed for evaluating the effect of this toxin on global transcriptional responses in Fg.
Phytohormones mediate immune responses in plants after pest or pathogen attack (Pieterse et al., 2009). In turn, plant pathogenic fungi have evolved ways to compromise host hormone pathways. This is achieved by degrading or producing phytohormones or interfering with their signaling pathways (Kazan and Lyons, 2014; Patkar and Naqvi, 2017). For instance, emerging evidence suggests that phytohormones such as abscisic acid (ABA), gibberellic acid (GA) and ethylene (ET) produced by fungi participate in pathogenicity (Chanclud and Morel, 2016). Previous studies indicated that Fg can likely produce auxin (IAA) and ET that may be utilized for attenuating host defenses during FHB (Foroud et al., 2019; Luo et al., 2016; Svoboda et al., 2019). In addition, SA hydroxylases were proposed to be involved in the degradation of host SA by Fg (Hao et al., 2019; Qi et al., 2019; Rocheleau et al., 2019). However, how hormonal compounds produced by Fg or the host plant are metabolized or involved in host infection is poorly studied. This is at least in part due to potential co-existence of phytohormones derived from both host and the pathogen in the infected tissue and the lack of knowledge on fungal genes involved in phytohormone biosynthesis.
Currently, potential molecular mechanisms employed by Fg during root infection are unknown (Kazan and Gardiner, 2018a). A global transcriptome analysis would provide a powerful way to broadly reveal previously unknown features during root infection, thus promoting development of new strategies for combating this pathogen. In addition, to what extent DON may affect global transcription in the pathogen has not been investigated. We previously reported an RNA-seq based transcriptome profiling of Fg prior to its physical contact with Bd roots. This analysis enabled us to discover novel genes that are involved in nitric oxide (NO) production in Fg upon sensing of root signals (e.g. metabolites found in the root exudates) and pathogen virulence (Ding et al., 2020). In this study, using Bd as a cereal model, we asked how Fg behaves as a root pathogen. To answer this question, we analyzed the transcriptome of Fg during infection of Bd roots. We analyzed phytohormone levels of infected Bd roots to understand potential roles played by phytohormones derived from the fungus. In addition, by comparing the transcriptomes of WT Fg and the DON deficient Tri5 mutant, we uncovered novel insights into global effects of DON on fungal gene expression and metabolism during the infection of host roots.
Materials and methods
Plant and fungal materials and root infection assay
The Fg CS3005 WT, Tri5 mutant (Desmond et al., 2008) and Tri5-GFP expressing (Gardiner et al., 2009) strains were routinely maintained on Potato Dextrose Agar (PDA, BD Difco). Bd (Bd21-3) seeds were surface sterilized and pre-germinated on filter paper (Whatman) placed in 150mm x 25 mm petri dishes (Corning) for 5 days. Bd roots inoculation with the Fg strains was carried out as described previously (Ding et al., 2020). Briefly, agar plugs (0.25 cm diameter) taken from fungal culture plates (Carboxymethylcellulose agar) were transferred to center of minimum media (pH 7) plates and pre-grown for 3 days. Five-day-old Bd seedlings were placed above the fungal colonies and inoculated for additional 5 days. Three biological replicates for Fg WT and ΔTri5 inoculated seedlings, and the Fg WT alone mycelia were produced by pooling materials from 10-12 plants or fungal mycelia.
RNAseq and transcriptomic analyses
Fungal and root materials were frozen in liquid nitrogen immediately after harvest. Total RNA was extracted from homogenized samples using a Qiagen RNeasy plant RNA extraction kit with on-column DNase I (Qiagen) digestion following manufacturer’s instructions. RNA was quantified and quality-checked prior to sequencing. An Illumina HiSeq2500 High Output platform was used to generate 50-base pair single-end reads (Australian Genome Research Facility). Reads quality control, alignment, transcript abundance and differential expression (DE) analyses were performed according to the method described previously (Ding et al., 2020). For pairwise comparison (FgWT-only vs. FgWT-Bd, or FgWT-Bd vs. ΔTri5-Bd), different sample files were normalized and merged. Reads were measured as FPKM (Fragments Per Kilobase of gene model per Million reads mapped), and a normalization method developed by Hart et al. (2013) was used and to eliminate background noise of FPKM values where genes with a Gaussian-fit derived log2(FPKM) value higher than -3 were considered as expressed. |log2 fold change| ≥ 1 and Benjamini and Hochberg-adjusted P value < 0.05 were applied to DE genes. RNAseq data are available at NCBI under the accession no. PRJNA631873.
Annotation and functional categorization of differentially expressed genes (DEGs)
BLAST2GO (Götz et al., 2008) was used to assign annotations for fungal DEGs. BLASTP reciprocal best hit analyses were performed in order to identify putative orthologous genes and match unique gene identifiers of the Fg CS3005 and PH-1 strains (Gardiner et al., 2014). Based on the PH-1 identifiers, classification ontology of DEGs from pairwise comparisons was annotated with FungiFun2 and subsequently subjected to enrichment analyses (https://elbe.hki-jena.de/fungifun). Functional categories were considered as enriched in the genome if an enrichment Benjamini-Hochberg adjusted P value is smaller than 0.05. Prediction of protein cellular localizations, secretome and putative effectors, transporters, carbohydrate-active enzymes, lipases, secondary metabolism enzymes and transcription factors was according to previously described methods (Ding et al., 2020).
Identification of homologous genes between Fg and Fp
All RNAseq reads were mapped to the Bd reference genome first. Unmapped reads were extracted and aligned to the Fg CS3005 and Fp CS3096 reference genomes, respectively, as per previously described (Ding et al., 2020). Only reads that could be mapped to both genomes were retained. Next, read counts measured as FPKM values were log transformed, normalized and subjected to PCA analysis (Fig. S4). Homologous genes within Fg and Fp were identified using a reciprocal best BLAST hit (RBBH) approach. Gene orthologs with identity of equal or higher than 99% were kept and used for syntenic analysis using a R package shinyCircos (Yu et al., 2018).
cDNA synthesis and quantitative real-time PCR analysis
0.5-1 µg total RNA was prepared for first-strand cDNA synthesis using the superscript IV synthesis kit (Invitrogen, USA) and quantitative real-time RT-PCR (qRT-PCR) was performed using the ViiA 7 real-time PCR detection platform (Applied BioSystems). Primers were based on previous studies (Ding et al., 2020; Voigt et al., 2005). Expression levels were normalized to the fungal house-keeping gene α-Tubulin and were averaged over three biological replicates.
Root sectioning and fluorescence microscopy
Root dissection was carried out following a previous described method (Ursache et al., 2018). Briefly, roots were harvested at 5dpi and fixed with 4% paraformaldehyde (Sigma) in PBS buffer (pH=6.9) overnight, washed twice with PBS buffer and cleared with ClearSee solution (Ursache et al., 2018) for 3 days. After clearing, roots were hand sectioned from 3 cm above the tips and stained with 0.2% Basic Fuchsin (Sigma). To image GFP and Basic Fuchsin fluorescence, root samples were observed using 488-nm excitation / 519-nm emission, and 561-nm excitation / 625-nm emission, respectively, on a Zeiss Axio Imager M2 microscopy.
UHPLC quantification of metabolites
For metabolite extractions, mock and Fg-infected roots, fungal mycelia and media samples were collected at 5 dpi, immediately frozen and ground in liquid nitrogen. 100 mg of fine-ground materials were resuspended in 2 mL extraction buffer (Ethyl acetate: methanol: dichloromethane, 3:2:1 v/v). 5 μL extracts from 6 replicates of each conditioned sample were injected into a Waters ACQUITY ARC UPLC system with Photodiode Array and passed through a Phenomenex Kinetex column (C18, 1.7 μm, 100 × 2.1 mm). The mobile phases consisted of solvent A (10 mM ammonium formate in water) and solvent B (10 mM ammonium formate in acetonitrile). The gradient program was a linear gradient from 2-40% solvent B delivered over 22 min followed by 40-80% B over 1.5 min at a constant flow rate of 0.4 mL per minute. As external standards, jasmonic acid (Sigma, 10 mM), methyl-jasmonate (Sigma, 40 mM), salicylic acid (Sigma, 10 mM), deoxynivalenol (Sigma, 100 ng/μL), gibberellic acid (Sigma, 10 mM), indole-3-actic acid (Sigma, 2 mM) were freshly prepared for serial dilutions (10x, 100x, 1000x, 10000x). Metabolites were quantified according to the concentration-gradient derived standard curves of external standards. Chromatography absorbance data were aligned using the following wavelengths: 204nm (for SA, JA and MeJA), 254nm (for IAA and DON) and 214nm (for GA), and extracted using the Empower 3 software (Waters).
Results and discussion
The Fg transcriptome during Bd root infection
Despite various studies investigating the transcriptome of Fg during colonization of above-ground tissues, Fg transcriptome during root infection has not been studied before (Kazan and Gardiner, 2018a). Therefore, here, we first investigated the Fg transcriptome during root infection of the model host Bd. This transcriptome experiment, which was initially designed for the discovery of a novel host-sensing mechanism in Fg prior its physical contact with roots, included three replicates of each of 1) Fg grown on minimal media (MM) (Fg-only) 2) FgWT and Bd roots grown together on MM without physical contact, 3) WT Fg colonizing Bd roots on MM (FgWT colonization) and 4) a Tri5 mutant colonizing Bd roots (ΔTri5 colonization). Previously, by comparing 1 with 2, we discovered new regulators involved in host sensing-mediated NO production (Ding et al., 2020). Here, we report on the comparisons between 1 and 3, 1 and 4 and 3 and 4, as detailed analyses of these have not been reported previously.
A total of 174,399,800 single-end reads, including 114,577,185 reads from Fg alone and Fg-infected Bd root samples (Ding et al., 2020) and 59,822,615 reads from the ΔTri5 root samples were generated by Illumina sequencing of mRNA libraries (Suppl. data 1). Prior to read mapping to the Fg CS3005 genome (Gardiner et al., 2014), reads aligned to the Bd genome were filtered. Of these filtered reads across all the samples, at least 91%, ranging from 5 to 19.9 million, were mapped to the Fg reference genome. Among these mapped reads, less than 0.3% matched to multiple genomic locations (Suppl. data 1). Out of the 12590 transcripts detected, 11942, 11291 and 11327 were found actively expressed in Fg- only, in WT Fg colonization and in Tri5 mutant colonization conditions, respectively.
Our analysis revealed a total of 2049 genes that were differentially expressed (DE) (log2 FC ≥ 1 and adjusted p value < 0.05) (Fig. 1A, Suppl. data 2) during colonization of Bd roots relative to Fg-only. Of these, 1281 and 768 were up- and down-regulated, respectively, during root colonization. The proportion of DE genes (DEGs) (around 20% of total number of genes found in the Fg genome) was similar to those observed during above-ground infection by Fg in other studies (Brown et al., 2017; Puri et al., 2016). A previous study indicated that some defense genes are similarly regulated in wheat roots and spikes in response to Fg (Q. Wang et al., 2018), suggesting that common strategies might be employed by Fg to infect different tissue types. Indeed, we previously reported that fungal knockouts of several DEGs identified here showed defects in both root and head infection (Ding et al., 2020).
Fungal processes employed by Fg during root infection
Detailed analysis of DEGs (Fg-only vs WT Fg colonization) revealed a number of enriched functional categories likely to be used by the pathogen during root infection. Below, some of these functional categories were discussed in more detail.
Genes encoding plant cell wall degrading enzymes (CWDE)
The Fg genome comprises a large number of genes encoding hydrolytic enzymes, transporters, secreted proteins and multiple gene clusters associated with secondary metabolite biosynthesis (Scherm et al., 2013; Sieber et al., 2014). These enzymes, non-enzymatic proteins and secondary metabolites together are generally considered fungal pathogenicity factors and their deployment during Bd root colonization indicates the importance of these pathway for sustaining the infection process. Indeed, over 30% of DEGs encode secondary metabolism enzymes (SMEs), secreted proteins, carbohydrate-active enzymes (CAZymes) and transporters (Fig. 1A). The most enriched categories for significantly up-regulated genes were associated with carbohydrate hydrolytic pathways (Suppl. data 3). Functional annotations showed that over half of the in planta activated CAZymes displayed modular structures related to CWDE acting on wall polymers such as cellulose, hemicellulose, lignin and pectin (Fig. 1B, Suppl. data 4), indicating the utilization of carbon from plant cell walls is a prominent capability of Fg during the colonization of Bd roots.
Genes encoding secreted proteins and putative effectors
Fungal pathogens produce many small secreted proteins or effectors to help facilitate host colonization. However, relatively little is known about potential Fg effectors. By comparing our DEGs with the previously defined Fg secretome (Brown et al., 2012) and using the EffectorP 2.0 software (Sperschneider et al., 2018), we identified 250 putative secreted proteins along with 65 predicted fungal effector encoding genes (Fig S1C, Suppl. data 4 and 5). Many of these secreted protein genes, including 189 induced ones (Suppl. data 5), encode putative lipases and peptidases predicted to perform hydrolytic functions (Fig. 1C). Among the 65 predicted effector-encoding genes, 52 were significantly up-regulated during Bd root infection (Suppl. data 5). Some of these DE genes were reported to encode effectors actively secreted by Fg during in vitro growth. For example, FG05_04074 encodes a protein of unknown function detected in two different secretomes (Lu and Edwards, 2016; Yang et al., 2012). Other putative effectors with annotated functional domains also found in previous studies included two glycoside hydrolases (FG05_11037 and FG05_06466), a putative acetylesterase (FG05_11280), a putative endonuclease (FG05_03365) and a cerato-platanin family protein (FG05_10212) for which roles of protection against host defense have been proposed (Lu and Edwards, 2016; Quarantin et al., 2016; Yang et al., 2012). Interestingly, the top induced effector candidates were mostly with predicted enzymatic functions (Suppl. data 5). Function of these differentially regulated putative effectors, such as FG05_04735 encoding a putative hypersensitive response-inducing elicitor and FG05_02255 a LysM domain containing protein (Suppl. data 5), could be predicted in comparison with those containing similar structural domains from different fungal pathogens and generally associated with host penetration, spore dispersal, triggering plant defense responses, inhibiting chitin-induced immunity or protecting against plant lysis (De Jonge et al., 2010; Khan et al., 2016; Lo Presti et al., 2015; Marshall et al., 2011; Mentlak et al., 2012). Although exact functions of these genes up-regulated during infection are largely unknown, it can be speculated that these secreted proteins and putative effectors could benefit the fungus during the colonization of host roots.
Genes encoding secondary metabolism enzymes (SME)
Fg is known to produce many secondary metabolites (SMs) during infection (Ma et al., 2013). In line with this, a strong induction of expression could be observed for genes encoding key signature enzymes (Fig. S1A), including the longiborneol synthase CLM1 (FG05_10397), butenolide synthase (FG05_08079), TRI5 and the terpenoid synthase DTC1 (FG05_03066) during infection (Suppl. data 6), suggesting that the corresponding products culmorin, butenolide, trichothecene and carotenoid may be the major mycotoxins delivered by the fungus to facilitate root infection. Among them, butenolide and trichothecene pathways are known to be co-regulated in vitro and in planta (Sieber et al., 2014). In contrast, down-regulation or very low in planta expression of other key SME genes, such as PKS12, NPS2, NPS1, PKS4 and PKS10 (Suppl. data 6), indicates that certain types of mycotoxins such as aurofusarin, ferricrocin, malonichrome, zearalenone and fusarin C might not be highly produced by Fg during root colonization. Aurofusarin does not affect wheat head infection by Fg (Malz et al., 2005), whereas ferricrocin and malonichrome have been shown to be important for pathogenesis-related development of Fg (Oide et al., 2014). The tailoring enzyme genes are usually clustered and co-regulated with the corresponding signature enzyme genes in Fg. These genes encode cytochrome P450s, oxidoreductases, acyltransferases and methyltransferases mainly involved in the SM pathway responsible for biosynthesis and modification of SM products (Sieber et al., 2014). Therefore, up-regulation of a large portion of tailoring enzyme genes found in this study is consistent with the regulation pattern of the signature enzymes (Fig. S1B).
Genes encoding fungal transporters
Transporter encoding genes mostly induced during Bd root colonization comprised a large group within the DE gene list (Fig. S1C, Suppl. data 7). Indeed, the regulation of transporter genes, particularly those associated with carbohydrate and nitrogen uptake as well as the ATP-binding cassette transporters (ABC transporters), is often linked to fungal nutrient assimilation, sensing, defense and pathogenicity status in pathogenic fungi (Abou Ammar et al., 2013; Coleman and Mylonakis, 2009; Divon and Fluhr, 2007; Gardiner et al., 2013; Schuler et al., 2015; Struck, 2015; Yin et al., 2018). Interestingly, the major facilitator superfamily (MFS) transporters associated with phosphate (Pi) transport and multidrug resistance (MDR) were mostly down-regulated during root infection compared to media alone controls (Fig. S1C, Suppl. data 7). During colonization of maize stalk, Fg overcomes Pi limitation by up-regulating high-affinity Pi transporter genes FGSG_03172 and FGSG_02426 (Zhang et al., 2016). The observation here that expressions of these genes during Bd root infection significantly reduced indicates a relatively rich root Pi environment under the experimental conditions and the time point examined. The down-regulation of MDR transporter genes suggested that they are possibly not essential for Fg to resist against root-derived anti-fungal compounds or self-derived toxins. The elevated expression of a large number of MFS-type carbohydrate transport genes together with the induction of PCWDE genes indicate that Fg preferentially utilizes carbon to accomplish the infection cycle and a state of glucose depletion may exist at the examined stage. The top induced ABC transporters (Suppl. data 7) exclusively belonging to the ABC-G type transporters are known to be associated with self-protection, possibly by effluxing of antifungal compounds in many pathogenic fungi (Coleman and Mylonakis, 2009). Thus, it is likely that these ABC transporters together with the MFS family multidrug resistance transporters could contribute to the virulence and fitness of Fg by detoxifying plant defense compounds.
We also observed that genes encoding amino acid-related transporters such as the amino acid/polyamine/organocation (APC) family and the amino acid/auxin permease (AAAP) family transporters, which are the major nitrogen transporters, were differentially expressed during root infection. In contrast, no inorganic nitrogen transporter gene showed altered expression (Suppl. data 7). This suggests that Fg root infection requires plant-derived organic nitrogen sources, and is consistent with the finding that polyamines as well as their amino acid precursors are potent DON inducers in Fg and play important roles during head infection (Gardiner et al., 2010, 2009). Interestingly, the highest induced transporter (FG05_02278, over 10-fold logFC) gene encodes a putative APC family protein transporter involved in choline uptake. Choline was identified as one of the major fungal growth stimulators in wheat anthers and implicated in promoting Fg virulence (Strange et al., 1972). Thus, it is possible that choline, in addition to amino acids and their derivatives, is another major factor contributing to Fg root colonization.
A small set of ‘core’ genes is activated during infection of different hosts and tissues by Fg
To obtain additional insights into Fg pathogenicity, we compared the Fg genes found to be induced during root colonization in this study with those previously reported to be induced during the colonization of other hosts or tissues (Brown et al., 2017; Harris et al., 2016; Lysøe et al., 2011; Zhang et al., 2012, 2016). These previous studies have reported several subsets of in-planta expressed Fg genes at multiple infection time-points and different disease development stages. To make a broader comparison, Fg genes that were commonly induced during infection at any of the studied time-points were selected. These included 3591 Fg genes induced during the infection of wheat heads (Brown et al 2017), 5061 genes expressed during the infection of wheat and barley heads and maize ears (Harris et al 2016) as well as 344 and 3066 genes induced during the infection of wheat juvenile coleoptiles (Zhang et al., 2012) and maize stalks (Zhang et al., 2016), respectively (Suppl. data 8). Through these comparisons, a total of 38 Fg genes commonly induced across all gene lists were identified (Fig. 2A, Suppl. data 9). Some of these genes were also differentially expressed between a FHB resistant and a susceptible wheat genotype (Pan et al., 2018). Unexpectedly, no mycotoxin- or pathogenicity-related SME genes were present among these 38 genes (Fig. 2B). An ABC transporter gene, FG05_04580 (FgABC1), and its flanking neighbor FG05_04581, which encodes a transcription factor highly inducible by the mycotoxin zearalenone (Lee et al., 2010), were found among these common genes. Deletion of FgABC1 causes reduced virulence of Fg on tested wheat tissues (Abou Ammar et al., 2013; Gardiner et al., 2013). Interestingly, FgABC1 and FG05_04581 homologs in the closely related pathogen F. culmorum were both highly induced by the antifungal compound tebuconazole (Hellin et al., 2018). This indicates that FgABC1 and FG05_04581 could be involved in self-protection against various defensive chemicals consistent with the observation that FgABC1 contributes to protection against the fungicide benalaxyl (Gardiner et al., 2013). Furthermore, most of these common genes encode non-SM enzymes such as CAZymes, peptidases and putative effectors. Among them, FG05_03624, a gene encoding a secreted xylanase, was previously shown to promote necrosis during Fg head infection (Moscetti et al., 2015). Protein homologs of several of these genes were also shown to be virulence factors in other fungal pathogens. For instance, FG05_00028 is homologous to metallopeptidases (MEP1), which were shown to be apoplastic effectors in F. oxysporum and M. oryzae (Jashni et al., 2015; Yan and Talbot, 2016). In F. oxysporum, the metallopeptidase FoMEP1 and the serine protease FoSEP1 act synergistically to cleave host chitinases and prevent their degradation of fungal cell walls (Jashni et al., 2015). Indeed, a serine-type proteinase inhibitor encoded by FG05_08012 found in our gene list, shows high similarity to FoSEP1 (Suppl. data 9). Another putative hypersensitive inducing elicitor FG05_04741 shows significant homology to the Verticillium dahlia effector PevD1, which was shown to be a secreted elicitor triggering host defense and cell death (Liang et al., 2018). We hypothesize that these putative Fg effectors may perform roles that are similar to those found in other fungal pathogens. Taken together, it can be hypothesized that a common set of Fg genes seems to play essential roles in Fg for successful colonization of different tissue types.
Fg genes specifically upregulated in Bd roots
We have identified 257 Fg genes that were exclusively upregulated during Bd root infections (Fig. 2A). Functional category analysis showed a significant enrichment for genes involved in transmembrane transport and cellular import (FDR = 0.00334). Of 34 transporter encoding genes induced, 12 were predicted to be associated with carbon transport (Suppl. data 10). This supports the finding discussed above that carbon utilization by Fg plays a role during Bd root colonization. Three ABC-G and two MDR transporters (Suppl. data 10) found among the enriched transporters might be specifically associated with detoxification of Bd root defense compounds. Among the four predicted effectors induced in roots (Suppl. data 10), the putative host-necrosis inducer protein FG05_10212 was shown to be constitutively expressed during infections of wheat heads and in vitro and confirmed as an extracellular protein (Lu and Edwards, 2016). The induction of this effector might contribute to necrosis observed in the infected Bd roots. Notably, of the seven putative Fg PCWDEs whose transcripts were only induced in Bd roots, five use lignin as substrate (Suppl. data 10), suggesting that lignin-degradation by Fg in Bd roots. Overall, while some common infection strategies may be employed by Fg during infection of different hosts and tissue types, there appears to be also unique processes used based on the activation of specific Fg genes during root colonization.
Partially shared infection strategies may be used by Fg and its sister species Fp during above- and below-ground infection of Bd
Previously, above-ground responses to the infection of Bd seedlings by F. pseudograminearum (Fp), another fungal species that is highly similar to Fg at the whole genome level (Gardiner et al., 2018) and was previously considered to be the same species as Fg (Kazan and Gardiner, 2018b), have been investigated (Powell et al., 2017). Both Fg and Fp show highly similar infection patterns on Bd (Fitzgerald et al., 2015). In addition, most genes are located in similar genomic regions in both fungi, whereas only a few species-specific genes, which could not be revealed by syntenic analysis, were found in genomic locations displaying high SNP densities (Gardiner et al., 2018). To further explore organ specificity of Fg infection on the same host, we compared the transcriptome of Fg with that of Fp (NCBI accession no. SRR3695327), during above ground infection of Bd at the same time point. Only Fg and Fp orthologous genes, which could be mapped to both of the genomes and share an identity of ≥99% were retained in this comparison. This stringent cut-off allows comparisons of only highly conserved genes that might be predicted to show similar pattern in expression and function in these two closely related fungal species.
In total, 1835 of the Fg DEGs were matched to Fp and formed a syntenic map (Fig. 3A and Suppl. data 11) consistent with the previously revealed genome structures (Gardiner et al., 2018). These genes were, in general, similarly expressed in Fg and Fp as indicated by the logarithmic transformed FPKM values (Fig. 3A). However, some variability in gene expression between the two species could be observed, particularly for genes found on chromosome 2 where the Fg orthologs tend to be preferentially expressed, as reflected by the expression heat map (Fig. 3A). Parts of chromosome 2 were previously identified as regions of the Fg genome that are rapidly evolving (Sperschneider et al., 2015). By manual curation, we selected and annotated the top 20 variant genes (Fig. 3B) that included three MFS-type transporter genes and an acetate permease homolog as well as several putative defense associated genes encoding a cell-wall glycoprotein (FG05_03352), peptidases (FG05_08075 and FG05_08141), and glucosidases (FG05_03387 and FG05_08265). A Zn2Cys6 transcription factor (FG05_03727), which shares the highest similarity to the yeast multidrug and oxidative stress resistance regulator STB5 (Larochelle et al., 2006), was identified. Accordingly, we found several oxidative stress responsive genes encoding a molybdopterin oxidoreductase (FG05_02880), NADH-flavin oxidoreductase (FG05_08077) and a putative flavohemoglobin (FG05_04458). Notably, FG05_03914 encoding a putative isochorismatase (ISC) gene was only expressed in Fg during infection of Bd roots. ISC-like effectors in filamentous pathogens are conserved virulence factors that can subvert plant salicylic acid (SA) pathway and interfere with host immunity (Liu et al., 2014). Resistance to biotrophic and hemi-biotrophic pathogens is usually conferred by the host SA pathway (Pieterse et al., 2012), and therefore, ISCs might be employed by Fg to support its hemi-biotrophic lifestyle to attenuate SA produced by the roots. The absence of Fp ISC transcription is consistent with the observation that host SA levels were not elevated in Bd plants at the early stages of infection (Powell et al., 2017). Host plants may activate tissue-specific defense signaling in response to below-and above-ground attacks (Lyons et al., 2015). To manipulate such defense responses, a fungal pathogen must evolve to a high flexibility for successful infection progressed in different host tissues. Despite the use of two fungal species, the data provided here may suggest shared infection strategies, including the interference with host defense, employed by Fg to support its belowground colonization.
DON influences different fungal processes in Fg during Bd root infection
The trichothecene mycotoxin DON has been shown to significantly inhibit Bd root growth (Pasquet et al., 2016). However, to the best of our knowledge, there has not been any study examining the effect of DON on different fungal processes in Fg. Therefore, we conducted an RNA-seq analysis by infecting Bd roots with DON producing and deficient strains to determine Fg genes whose expressions are modulated by DON. We first focused on genes up-regulated in the DON deficient ΔTri5 mutant strain relative to WT. Of 973 genes differentially expressed between ΔTri5 and WT, 432 genes were expressed at higher levels in ΔTri5 (Fig. 4A, Suppl. data 12). These genes were subjected to functional enrichment analysis based on the MIPS FGDB (Fusarium graminearum Genome Database Functional Catalogue classification) (Güldener et al., 2006). This analysis showed that these genes are enriched for transport (FDR = 0.0016) of carbon-compounds, carbohydrates, and heavy metal ions, disease, virulence and defense (FDR = 0.005) and homeostasis of phosphate (FDR = 0.02). A relatively smaller portion of fungal pathogenicity and metabolism associated genes encoding CAZymes, SMEs, transporters and secreted proteins were induced in ΔTri5 relative to WT (Fig. 4B-C, Fig. S2 and suppl. data 12-13). Only metabolic pathway genes encoding methyl- and glycol-transferases as well as phosphate, lipid and polyamine transporters were preferentially induced (Fig. S2). Host-derived phosphates, lipids as well as polyamines may influence Fg infection in multiple host tissue types (Gardiner et al., 2010; Zhang et al., 2016). Their enhancement might be due to a positive feedback to the lack of DON to balance the fungal metabolism and cell structure in root proliferation. In addition, 28 CAZyme encoding genes were expressed higher in ΔTri5 than in WT and many of these were putative PCWDEs involved in lignin degradation (Fig. 4B). Lignin is one of the major barriers against fungal pathogens (Bhuiyan et al., 2009). Defense related or unrelated lignin content at fungal penetration sites might affect Fg intra-cellular progression (Zhang et al., 2016). In wheat roots, Fg colonization could be observed in lignin-rich vascular bundles (Bhandari et al., 2018; Wang et al., 2015). Therefore, the upregulation of lignin-degrading enzymes in ΔTri5 could be beneficial to the fungus to compensate DON deficiency and assist root colonization. In line with this, we observed reduced lignin deposition in roots colonized by ΔTri5 relative to the roots either mock-inoculated or colonized by WT Fg (Fig. 5).
Seven of the 33 secreted protein genes induced in ΔTri5, including a pathogenesis-related protein 1 (PR1) homolog (FG05_03109) and a putative cutinase (FG05_03457), may be considered putative effectors. The most differentially regulated SME genes in ΔTri5 were tailoring enzyme genes encoding cytochrome P450s and oxidoreductases (Fig. S2A and S2B). Surprisingly, most DEGs with elevated transcripts levels in ΔTri5 during root colonization (405 out of 432 genes) were expressed either significantly lower in WT during Bd root infection than Fg only or remained unchanged comparing to WT in vitro (Suppl. data 12). Besides, we noticed that none of the above-mentioned core genes was reduced in ΔTri5 during infection, supporting the notion that these genes may contribute to infection more than others, independently of fungal DON production.
We next looked at the 541 significantly downregulated genes in ΔTri5 relative to WT in roots. The most enriched functional categories during root infection were C-compound and carbohydrate metabolism (80 genes, FDR = 0.00008), disease, virulence and defense (16 genes, FDR = 0.01), secondary metabolism (27 genes, FDR = 0.01), protein or peptide degradation (26 genes, FDR = 0.01) and transport facilities (42 genes, FDR = 0.03) (Suppl. data 14). In addition, we found two sets of adjacent genes FG05_02297-FG05_02309 and FG05_08077-FG05_08084 that showed reduced expression in ΔTri5. Of these two sets, the latter genes, which are part of the mycotoxin butanolide cluster, shared similar regulation pattern with the Tri (tricothecene) cluster genes during infection of wheat heads (Boedi et al., 2016). In the saprophytic fungus Trichoderma arundinaceum, the loss of trichothecene production likely contributed to an increase of fungal secondary metabolites (Lindo et al., 2019, 2018). Therefore, DON seems to affect fungal metabolism during Fg infection in Bd roots.
Previously, Tri5 deletion was reported to lead to observable metabolic changes in Fg growing in rich medium, suggesting that DON might be linked to fungal physiology and development (Chen et al. 2011). In line with this, we also found that several differentially regulated TF genes in ΔTri5 during root infection were putative development-associated regulatory genes (Suppl. data 15). For example, FG05_08892 (MAT1-1-1) and FG05_05151 are known to be associated with sexual development (Kim et al., 2015), FG05_01139 (FgCBF1) is a predicted chromatin remodeling regulator (Guo et al., 2016), and FG05_03597 is homologous to Aspergillus nidulans FlbA, which is required for the control of mycelial proliferation and activation of asexual sporulation (Yu et al., 1996). However, when grown on MM, Tri5 was barely expressed in vitro, and no DON could be detected by metabolic analysis (Fig. 6D). ΔTri5 also did not show any growth defects (Chen et al., 2011). Together, our results suggest that ΔTri5 may colonize the roots by utilizing a small set of genes not used by the WT fungus.
DON is produced during Bd root colonization but does not act as a virulence factor
DON is a virulence factor during infection of wheat heads by Fg (Wang et al., 2020). However, it is unknown if this mycotoxin could also act as a virulence factor during infection of Bd roots. To determine this, the infection process was monitored using a Fg strain expressing Tri5-GFP fusion driving by the native Tri5 gene promoter (Gardiner et al., 2009). Strong GFP signals could be visualized two days post-inoculation (dpi) in inoculated roots, indicating that the infection was progressing, and DON production was initiated (Fig. 7A and 7B). Consistent with this observation, Tri5 was highly induced at 2 dpi and remained at high levels at later time points (3, 5, and 7 dpi) in the WT isolate (Fig. 7C). Previously, a temporarily similar infection pattern by Fg was also observed in wheat seedling roots (Wang et al., 2015). Fg root infection of wheat triggers induction of systemic defense responses in above-ground parts of the plant (Wang et al., 2018). In addition, DON preferentially inhibits root growth in wheat, Bd and Arabidopsis plants (Gatti et al., 2019; Masuda et al., 2007; Pasquet et al., 2016), and has been proposed to act as a major virulence factor in the early stages of Fg root infection (Wang et al. 2018). To assess the role of DON during root infection, a Fg Tri5 mutant was used in inoculation experiments together with WT and Tri5-GFP strains. Bd roots infected by ΔTri5 exhibited levels of lesion development that were like those caused by WT and Tri5-GFP at 7 dpi (Fig. 7D). Therefore, while DON is a virulence factor in Fg during FHB of wheat and is highly induced in roots, Bd root infection by Fg seems to be independent of DON production. Indeed, various phytopathogenic phenotypes have been described for Fg DON deficient mutants (Boenisch and Schäfer, 2011; Cuzick et al., 2008; Jansen et al., 2005). For instance, altered levels of DON have been shown to inhibit plant apoptosis-like programmed cell death (PCD) induced by heat stress in Arabidopsis (Diamond et al., 2013).
Phytohormone dynamics during Fg colonization of Bd roots
It is becoming increasingly evident that plant pathogens interfere with phytohormone pathways by producing plant hormones (Kazan and Lyons, 2014). However, pathogen-produced phytohormones have rarely been examined during root infections. We therefore next examined the transcriptome of Fg during Bd root infection, coupled with metabolic analyses, to determine putative phytohormone associated genes in Fg and their potential involvement in Bd root colonization.
JA produced by Fg is not associated with Bd root colonization
The oxylipin hormone jasmonic acid and its derived metabolites collectively known as jasmonates (JAs) are derived from lipid peroxidation and can affect both host and fungal physiological processes (Tsitsigiannis and Keller, 2007). Fungal oxylipin biosynthesis is catalyzed by lipoxygenases (LOXs) (Fischer and Keller, 2016). In F. oxysporum, FoxLOX was found to exhibit a multifunctional activity in oxylipins production, thus proposed to possess a function in JA pathways (Brodhun et al., 2013). Interestingly, we noticed that the Fg homolog of FoxLOX, FG05_05046, was expressed during in vitro growth, and highly induced during root infection (4.8 fold, Table S1). JA-regulated defenses in plants can be interrupted by pathogen-derived hormone analogs (Caarls et al., 2017; Patkar and Naqvi, 2017). The Fg genome does not contain a homolog of the M. oryzae antibiotic biosynthesis monooxygenase (Abm), which converts host-derived JA into 12-hydroxyjasmonic acid (12OH-JA), thus attenuating rice blast disease resistance (Patkar and Naqvi, 2017). The Arabidopsis 2OG oxygenases (JOXs) are responsible for JA hydroxylation (Caarls et al., 2017; Smirnova et al., 2017). We identified ten homologous of Arabidopsis JOXs in Fg (Table S1). Of these, FG05_08081 and FG05_02301, whose protein products share 22-26% identity to JOXs, were induced by 8.3 and 3.9 fold, respectively, during root infection. FG05_08081 is present in the butanolide biosynthesis gene cluster, members of which were also significantly upregulated during root infection (Suppl. data 2). However, whether FG05_08081 and FG05_02301 are involved in JA degradation requires further analyses.
JA-associated host defense against Fg has been studied during FHB development. Inhibition of JA by DON at the bottom of wheat florets promotes fungal progression through rachis notes (Bönnighausen et al., 2019). Furthermore, late activation of JA signaling during FHB has been proposed to correlate with a necrotrophic transition of Fg (Ding et al., 2011). To determine if JA levels change during root infection, we quantified JA levels in Bd roots either mock-treated or inoculated with WT Fg or ΔTri5 strains by high performance liquid chromatography (HPLC). We found that 5 ng/mg dried material of JA was produced in Fg mycelia grown on MM (Fig. 6A). However, JA levels found in infected and control roots were much lower than 5 ng/mg and did not display any significant difference (Fig. 6A). Interestingly, however, higher levels of JA derivative methyl-JA (MeJA) were detectable in the roots infected by ΔTri5 than those infected by WT (Fig. 6B). Thus, DON seems to inhibit MeJA production in the infected Bd roots. This is consistent with the observation in wheat heads where MeJA levels in the ΔTri5-infected tissue were less than those infected with WT Fg (Bönnighausen et al., 2019).
SA may synergistically interact with Bd root defense-related metabolic pathways
SA is a major defense hormone typically associated with plant defense against biotrophic pathogens (Glazebrook, 2005). JA and SA accumulate at different basal levels in various wheat cultivars and antagonistically fine-tune host defense responses (Powell et al., 2017). Therefore, we next focused on Fg responses to SA. Although the SA-pathway may be involved in host basal resistance against Fg (Ding et al., 2011; Makandar et al., 2010), SA-associated systemic acquired resistance (SAR) played no role in FHB resistance (Li and Yen, 2008). During infection of Bd seedlings by Fp, SA biosynthesis was induced (Powell et al., 2017). However, the function of SA during root infection by Fg remains elusive.
In Arabidopsis, SA biosynthesis during pathogen infection mainly relies on the intermediate chorismate processed by isochorismate synthase I (ICS1), the GH3 acyl adenylase-family enzyme PBS3, and the BAHD acyltransferase-family protein EPS1 (Torrens-Spence et al., 2019). The phenylalanine ammonia lyase (PAL) pathway also mediates SA synthesis through the conversion of benzoic acid or coumaric acid, but only contributes to a small portion of total SA production (Wildermuth et al., 2002). Similarly, some bacteria can directly convert isochorismate to SA by isochorismate pyruvate lyase (IPL) (Serino et al., 1995). While the host is believed to be the source of SA production in various plant-bacteria interactions, whether there is a fungal origin of SA remains unknown. In the Fg genome, we found two ICS1 homologs FG05_05195 and FG05_12934. Of these, FG05_05195 was lowly expressed (FPKM < 0.5), but FG05_12934 exhibited constitutive and high transcript levels during both in vitro growth and infection of Bd roots (Table S1). Fg also has an EPS1 homolog, FG05_00237, which was significantly upregulated during Bd root infection (Table S1). FG05_09331 encodes a protein sharing about 50% identity with PAL. FG05_09331 expression decreased by 2.2-fold in infected roots relative to Fg grown in vitro. No PBS3 and IPL homologs could be identified in Fg. While expression patterns of ICS1 and EPS1 homologs may coincide with observed SA production by Fg, the absence of a PBS3 homolog indicates other components or pathways could be involved in the SA biosynthesis. To determine if SA is produced in Fg-Bd interactions, we extracted metabolites from Bd roots with or without Fg inoculation and quantified SA levels by HPLC. In Fg mycelia, there was also about 0.01 µg SA per mg dry material (Fig. 6C). Most SA was found in the roots inoculated with WT Fg, followed by uninfected roots and the lowest levels in the ΔTri5 infected roots (Fig. 6C), indicating a potential role for DON in regulating SA levels during root infection by Fg although it is difficult to estimate the exact contribution of Bd or Fg to the SA levels measured.
SA has a direct effect on Fg growth, most likely associated with active degradation of SA by fungal hydroxylases. In recent studies (Hao et al., 2019; Qi et al., 2019; Rocheleau et al., 2019), two proteins FGSG_08116 and FGSG_03657 have been characterized with a function in SA degradation. While transcription of both FGSG_08116 and FGSG_03657 can be induced by external SA, fungal virulence in wheat heads was only influenced by deletion of the former. The function of FGSG_03657 for SA degradation was not disabled in deletion mutants. Interestingly, FG05_03657 (a.k.a. FGSG_03657) was exclusively induced during Bd root infection (Suppl. data 10), suggesting that it may have a role in regulating SA levels in infected roots. To determine if Fg could possess additional putative SA hydroxylase genes (Hao et al., 2019), we searched the Fg genome and identified 28 homologs of the SA sensor and degradation protein Shy1 from Ustilago maydis with FgShyC displaying at least 20% identity to Shy1 (e-value < 10-5) (Fig. S3). This similarity is much higher than the values reported previously (Hao et al., 2019; Rabe et al., 2013), where the expansion of these putative proteins in Fg was supported by phylogenetic analysis (Fig. 8A). In our current transcriptome, FG05_03657 and additional ten genes encoding putative SA hydroxylases were significantly upregulated in Fg during root infection (Fig. 8B). Therefore, SA biosynthesis by the pathogen as well as the host can contribute to the expression and regulation of these genes.
Expression of genes involved in the biosynthesis of other phytohormones in Fg
during Bd roots infection
In addition to JA and SA, other phytohormones such as gibberellins (GAs), auxins (IAAs), ethylene (ET), cytokinins (CKs) and abscisic acid (ABA) also participate in modulating host defense signaling (Pieterse et al., 2012). GAs can be synthesized by a number of Fusarium species but not by Fg due to the lack of a corresponding biosynthesis gene cluster (Cuomo et al., 2007). As a virulence factor, GA is restricted to the necrotrophic fungal pathogen F. fujikuroi (Wiemann et al., 2013) and is possibly involved in attenuating host JA signaling (Navarro et al., 2008). Under our inoculation conditions, GA was detected in the roots infected by WT Fg but not by the ΔTri5 mutant (Fig. 6F), indicating an endogenous GA production in Bd roots upon Fg infection and a potential positive effect of DON on root GA production. Similarly, GA accumulates in wheat heads infected by Fg (Bönnighausen et al., 2019), thus Fg seems to trigger a GA-dependent response in host roots that is similar to one observed in wheat florets (Buhrow et al., 2016). However, the association between DON and GA production requires further investigations.
Fungal genes encoding indole-3-acetaldehyde dehydrogenases (Iad) and tryptophan aminotransferases (IaaM) were thought to be responsible for IAA production (Reineke et al., 2008). A possible third pathway for auxin biosynthesis could be mediated by Fg genes homologous to YUCCA, a key enzyme involved in plant auxin biosynthesis (Mano and Nemoto, 2012). Interestingly, only one of the three Fg Iad gene homologs, FG05_02773, showed more than 4-fold induction in Fg during root infection as compared to Fg grown in vitro (Table S1). In contrast, two IaaM homologs and a YUCCA homolog were significantly downregulated in Fg inoculated roots (Table S1). Thus, the strong induction of FG05_02773 might coincide with the production of the auxin indole-3-acetic acid (IAA) by Fg, which could thereafter compromise the host auxin pathway. Fungal auxin biosynthesis plays a role in pathogenicity of several pathogens (Chanclud and Morel, 2016). In Fg, auxin was proposed to be a virulence factor (Svoboda et al., 2019). Indeed, Fg is able to synthesize IAA but also sensitive to exogenous application of IAA and its biosynthetic intermediates (Qi et al., 2016). However, Fg infection strongly inhibited IAA levels in roots infected by Fg WT or ΔTri5 (Fig. 6E). This is contradictory to the findings in wheat where auxin levels increased during FHB (Wang et al., 2018). Factors such as host tissue types and hormone antagonists (Kazan and Manners, 2009) might be responsible for such differences.
Fg can exploit host ET signaling during colonization of both dicotyledonous and monocotyledonous plants and is believed to be capable of producing ET to counteract host defense pathways (Chen et al., 2009). However, rather than ET forming enzymes (EFE), Fg was thought to use pathways incorporating 1-aminocyclopropane carboxylic acids (ACC) as precursors for ET biosynthesis (Svoboda et al., 2019). Although an enzymatic function for two of the five ACC enzymes encoded by the Fg genome could be confirmed, fungal mutants for these genes showed no defect in pathogenicity on wheat (Svoboda et al., 2019). We looked at the expression of all these five genes, including the three annotated ACC synthase genes, FG05_05184 (ACS1), FG05_07606 (ACS2), FG05_13587 (ACS3), and two ACC deaminase (ACD) genes, FG05_02678 and FG05_12669, but found no differential expression during Bd root infections (Table S1), suggesting that pathogen produced ET may not be involved in root infection.
Biosynthesis of fungal cytokinins, which can be mediated by either one or both of the fungal transfer RNA-isopentenyl transferases (tRNA-IPT) and the Lonely Guy (LOG) enzyme, is associated with host immunity and nutrient modulation and maintenance of hemi-biotrophic lifestyles during infection (Spallek et al., 2018). Unlike many other Fusarium species, the Fg genome contains only one tRNA-IPT gene homolog (FG05_09015) (Sørensen et al., 2018). This gene was not differentially regulated (Table S1) and only moderately expressed (FPKM < 10) during both root infections and in vitro growth on MM.
Exogenous ABA has no effects on disease development, Fg toxin production or defense hormone levels in Fg-challenged wheat heads, but can promote fungal hydrolase and cytoskeletal reorganization genes induced early during infection and increase wheat’s susceptibility to FHB (Buhrow et al., 2016). The elucidation of fungal genes responsible for ABA production in the necrotrophic pathogen Botrytis cinerea and a few others has led to the hypothesis that a conserved ABA biosynthesis pathway exists in fungi (Lievens et al., 2017). In B. cinerea, such pathway involves four clustered genes BcABA1-4 and a sesquiterpene cyclase gene BcSTC5 (Izquierdo-Bueno et al., 2018). Fungal ABA was shown to act as a virulence factor in M. oryzae, which also harbors a direct ABA biosynthesis pathway but lacks a BsABA3 ortholog (Spence et al., 2015). Similar to M. oryzae, we could identify homologs of only BcABA1, 2 and 4 by BLASTp in Fg. In the current root transcriptome, none of these genes was differentially regulated (Table S1). Overall, while fungal GAs and IAAs might be associated with Fg root infections, it is unlikely that ETs, CKs and ABAs are involved in Bd root infection by Fg.
Conclusions
The results presented here provide a detailed overview of root infection strategies employed by Fg, an important cereal fungal pathogen. The transcriptional regulation of pathogen metabolic pathways, virulence factors and signalling events during root infection show both unique and common features to those employed by Fg when infecting above-ground tissues. The mycotoxin DON, although not required for fungal virulence, produced during root infection appears to broadly affect various fungal processes and interplay with host responses. Expressions of several fungal stress and defence genes might help the pathogen to effectively deal with plant defence responses. In line with this, fungal JA, IAA and, in particular SA, seem to be used to interfere with root defenses. The findings presented in this paper will be useful for dissecting the mechanism of Fg belowground lifestyle and the development of novel plant protection strategies.
Conflict of interest
All authors declared no conflict of interest.
Acknowledgements
We thank Di Xiao and Dr. Jonathan Powell for technical assistance. Yi Ding was the recipient of a post-doctoral fellowship from the Commonwealth Scientific and Industrial Research Organization Research Office.
References
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.↵
- 11.↵
- 12.↵
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.↵
- 19.↵
- 20.↵
- 21.↵
- 22.↵
- 23.↵
- 24.↵
- 25.↵
- 26.↵
- 27.↵
- 28.↵
- 29.↵
- 30.↵
- 31.↵
- 32.↵
- 33.↵
- 34.↵
- 35.↵
- 36.
- 37.↵
- 38.↵
- 39.↵
- 40.↵
- 41.↵
- 42.↵
- 43.↵
- 44.↵
- 45.↵
- 46.↵
- 47.↵
- 48.↵
- 49.↵
- 50.↵
- 51.↵
- 52.↵
- 53.↵
- 54.↵
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.↵
- 60.↵
- 61.↵
- 62.↵
- 63.↵
- 64.↵
- 65.↵
- 66.↵
- 67.↵
- 68.↵
- 69.↵
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.↵
- 75.↵
- 76.↵
- 77.↵
- 78.↵
- 79.↵
- 80.↵
- 81.↵
- 82.↵
- 83.↵
- 84.↵
- 85.↵
- 86.↵
- 87.↵
- 88.↵
- 89.↵
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.↵
- 99.
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.↵
- 106.↵
- 107.↵
- 108.↵
- 109.↵
- 110.↵
- 111.↵
- 112.↵
- 113.↵
- 114.↵
- 115.↵
- 116.↵
- 117.↵
- 118.↵
- 119.↵
- 120.↵
- 121.
- 122.↵
- 123.↵
- 124.↵
- 125.↵
- 126.↵
- 127.↵
- 128.↵
- 129.↵
- 130.↵
- 131.↵
- 132.↵